The Tufts University study, published in July's Journal of Experimental Social Psychology, showed that a control group of 15 undergraduate students, both male and female, could accurately tell whether or not a man was gay, within 50 milliseconds, about 60 percent of the time when shown a photograph taken from an Internet personal ad or social networking site Facebook. The subjects overall were accurate between 55 and 70 percent of the time.

This particular study only used pictures of men "because they have a greater presence on the Internet than women," according to the study's lead author Nicholas Rule.

A past by Rule used facial features of CEOs to predict their companies' profits. Rule also plans to repeat the "gaydar" study using pictures of women.